Villa R et al.
The rules defining which small fraction of related DNA sequences can be selectively bound by a transcription factor are poorly understood. One of the most challenging tasks in DNA recognition is posed by dosage compensation systems that require the distinction between sex chromosomes and autosomes. In Drosophila melanogaster, the male-specific lethal dosage compensation complex (MSL-DCC) doubles the level of transcription from the single male X chromosome, but the nature of this selectivity is not known1. Previous efforts to identify X-chromosome-specific target sequences were unsuccessful as the identified MSL recognition elements lacked discriminative power2, 3. Therefore, additional determinants such as co-factors, chromatin features, RNA and chromosome conformation have been proposed to refine targeting further4. Here, using an in vitro genome-wide DNA binding assay, we show that recognition of the X chromosome is an intrinsic feature of the MSL-DCC. MSL2, the male-specific organizer of the complex, uses two distinct DNA interaction surfaces—the CXC and proline/basic-residue-rich domains—to identify complex DNA elements on the X chromosome. Specificity is provided by the CXC domain, which binds a novel motif defined by DNA sequence and shape. This motif characterizes a subclass of MSL2-binding sites, which we name PionX (pioneering sites on the X) as they appeared early during the recent evolution of an X chromosome in D. miranda and are the first chromosomal sites to be bound during de novo MSL-DCC assembly. Our data provide the first, to our knowledge, documented molecular mechanism through which the dosage compensation machinery distinguishes the X chromosome from an autosome. They highlight fundamental principles in the recognition of complex DNA elements by protein that will have a strong impact on many aspects of chromosome biology.